The Newsroom · guide · 21 August 2026
Non-Commodity Score
The Non-Commodity Score (NCS), originated by Laurelin Labs, measures whether content deserves an AI citation across four vectors: information gain, experiential evidence, telemetry and connectivity.
By Piers Butler
The Non-Commodity Score (NCS) is a content audit, originated by Laurelin Labs, that measures the one property AI search actually rewards: whether a page says anything the already-retrievable competition does not. It grades four vectors, information gain against the live retrieval corpus, first-hand experiential evidence, telemetry (concrete numbers and conditions instead of adjectives), and connectivity (verifiable entity and source wiring), and every qualitative judgment must return verbatim quoted evidence from the page. Commodity content, however polished, scores low, because an AI synthesiser has no reason to cite a page that restates what it already retrieved.
Why does commodity content fail in AI search?
An AI answer is assembled from retrieved passages, and the assembler cites the passages that contributed something. A page that competently summarises the same points as the ten pages retrieved beside it is, from the synthesiser's perspective, redundant: it gets retrieved, contributes nothing unique, and is passed over in the citation list. That is commodity content: accurate, fluent, and interchangeable with its competition. The economics are unforgiving, because generative tools have made commodity content effectively free to produce, which means its citation value has fallen to zero precisely as its volume has exploded. Google's own guidance on generative AI search says the quiet part plainly:
Creating content that people find unique, compelling, and useful will likely influence your website's presence in generative AI search in the long run more than any of the other suggestions.
Google Search Central, AI features and your website, Google.
The problem for content teams is that "unique, compelling and useful" is a judgment, and unmeasured judgments do not survive contact with publishing schedules. The NCS exists to turn that judgment into a measurement with evidence attached.
What are the four vectors of the Non-Commodity Score?
| Vector | Question | What scores high |
|---|---|---|
| Information gain | Does the page add anything beyond what the corpus already retrieves for its queries? | Original data, findings and framings absent from the competing retrieved passages |
| Experiential evidence | Does the page show first-hand experience, the first E of E-E-A-T? | First-hand actions and measurements, specific conditions, outcomes with numbers, no generic filler |
| Telemetry | Is the page dense in concrete, checkable specifics? | Real figures, dates, settings and named methods where a commodity page would use adjectives |
| Connectivity | Is the page wired to verifiable entities and sources? | Structured data referencing real entities, and outbound links to authoritative primary sources |
The vectors are deliberately complementary: information gain measures the page against its competition, experiential evidence and telemetry measure the substance of the page itself, and connectivity measures how checkable that substance is. A page can only fake one of them at a time, which is the point.
How does the audit measure each vector?
Information gain is the vector that needs the retrieval corpus, and it is measured the way a RAG system would: the page's passages and the top-ranking competing pages for its query cluster are embedded into the same vector space, and the score reflects what the page contributes beyond that retrieved set. This is why the full measurement only exists in the audit proper; the free NCS Lite tool on laurelinlabs.com scores the other three vectors and renormalises, labelling the result directional. The approach is described in depth in the companion article on RAG analysis.
Experiential evidence is graded by a language model judge against a fixed rubric, first-hand actions, specific conditions, outcomes with numbers, absence of filler, with one hard rule: the judge must return short verbatim quotes from the page that justify the grade, and if no experiential evidence exists it must say so with an empty evidence list rather than a hedged score. Telemetry and connectivity are measured deterministically from the page itself: the density of concrete figures and specifics in the text, and the entity references, structured data and authoritative outbound links in the markup. Every vector's finding therefore arrives with its evidence attached, quoted spans for the judged vector, counted and named artefacts for the measured ones, so an editor can verify any line of the report against the page in seconds.
What does the audit report actually give you?
Three things a generic content audit does not. A ranked fix list per page, ordered by which vector is costing the most, with the evidence for each finding quoted inline; passage-level detail, because the unit of AI retrieval is the passage, so the report identifies which specific sections carry the page's gain and which are commodity filler; and a portfolio view across the audited estate, which is where the strategic finding usually lives, most estates discover that a small minority of pages carry nearly all of their citable substance, and the production system, not the writers, is what has been optimised for volume over gain. For Newsroom, that portfolio view is the editorial planning input: it says where a new brief should add evidence rather than add another article.
What are the honest limits of the score?
Two, stated plainly because a measurement tool that overclaims is itself commodity content. First, calibration: NCS outputs are directional grades until enough scored outcomes exist to calibrate them against real citation behaviour, and uncalibrated results are labelled uncalibrated rather than dressed as verdicts. The methodology deliberately withholds pass or fail gates until the calibration data justifies them. Second, scope: the NCS measures whether content deserves citation, not whether it can be retrieved at all; a high-scoring page behind an indexation failure is still invisible. That upstream layer belongs to the technical gate, covered in the companion article on the AEO gate checks, and the two audits are designed to run in that order.
Frequently asked questions
What does a Non-Commodity Score actually tell you?
Whether a page says anything the already-retrievable competition does not. It grades four vectors: information gain against the live retrieval corpus, first-hand experiential evidence, telemetry (concrete numbers and conditions rather than adjectives), and connectivity (how the page is wired to verifiable entities and sources). A high score means an AI synthesiser has a reason to cite you rather than paraphrase around you.
Who created the NCS methodology?
The Non-Commodity Score was originated by Laurelin Labs as part of its AEO audit engine, and is used by Newsroom Studio in its editorial pipeline. A free Lite version runs at laurelinlabs.com, and the full audit adds the SERP-corpus information gain measurement.
Is the NCS a pass or fail test?
No. Scores are reported as directional grades with quoted evidence, and uncalibrated outputs are labelled as such. The methodology deliberately withholds hard pass or fail gates until calibration data justifies them, because a confident number without evidence is exactly the kind of commodity claim the score exists to catch.